The research is aimed at studying the representation and processing of spatially distributed information in the somatosensory system of the primate. The plan includes psychophysical and neurophysiological experiments and the development of computational relationships between the two classes of results. We seek continued support for neurophysiological studies of the hand region of the primary somesthetic cortex of the primate. A battery of stimuli that have been used previously in peripheral neurophysiological studies and in psychophysical studies (excepting iv) will be used in the cortical experiments. These stimuli are: (i) tetragonal dot patterns scanned across the skin; (ii) full sets of embossed letters scanned across the skin; (iii) periodic mechanical gratings used in a previous psychophysical task; (iv) a dense array of independently controlled probes, currently under construction at the Applied Physics Laboratory. Our specific aims fall into three categories. Neurophysiological. The spatially structured stimuli described above will be used to study the response properties of neurons in the hand regions of areas 3b, 1, and 2. In the first (current) phase of these investigations, the aim is to identify, for further study, those neuronal groups whose responses might provide the basis for tactual spatial discrimination. In the second phase the aim is to study these groups, identified by area and submodality class, more extensively using stimuli (ii), (iii) and (iv). The objective here is to reconstruct the population responses evoked by these stimuli and to analyze them with respect to the observed psychophysical behavior. Behavioral. Modifications in the behavioral design are aimed at making the tactual test stimuli behaviorally relevant during cortical recording and at defining spatial tactual discrimination capabilities in the monkey. Analytical. Studies have begun that are aimed at establishing a theoretical framework for the comparison of pattern identification behavior and the undelrying patterns of neural activity. The specific aim is to develop computational methods that are firmly grounded in statistical decision theory.